Understanding, Predicting, and Remediating Perturbations to Host-Microbe Interactions

Antibiotics can acutely and persistently compromise microbiome stability and resilience, enabling enhanced colonization of pathogenic microbes, and loss of host-beneficial functions from depleted commensals. We interrogate resistome, taxonomic, and functional dynamics of microbiota of humans, animals, soils, and built-environments via high-throughput functional metagenomic selections, whole metagenome and transcriptome sequencing, and isolate phenotyping and whole genome sequencing. We use these datasets to model and predict the impact of various biotic and abiotic perturbations on the transmission and evolutionary dynamics of diverse microbiomes and their resistomes across time and habitats, through application of generalized linear mixed models and machine learning algorithms. By integrating community-level analyses of microbiome perturbation and response, with mechanistic analyses of novel microbial resilience functions, we are improving the accuracy and sensitivity of genomics-based methods for diagnosing and treating pathogens and their resistance properties in the clinic and maintaining or rescuing healthy microbiome states. 

Speaker:Gautam Dantas, Washington University, St. Louis